Evaluation of the DNDCv.CAN model for simulating greenhouse gas emissions under crop rotations that include winter cover crops

Abstract

Context Process-based modelling studies can help inform conservation practices for mitigating soil surface CO2 and N2O fluxes. Aims We evaluated the ability of the DeNitrification-DeComposition (DNDC) model to predict field-measured soil surface CO2 and N2O emissions in crop rotations managed with cover crop (CC) and without cover crop (NC) under the 27-year no-till field experiment in South Dakota, USA. Methods Emissions were measured in a 2-year corn–soybean and a 4-year corn–soybean–oat–winter wheat rotation. The model was calibrated with 2-year NC treatment and evaluated against three treatments (2-year CC, 4-year NC and 4-year CC) during the growing season of corn (2017) and soybean (2018). Key results Across all treatments, the model simulated soil temperature (MBE, −0.73–0.29°C; RMSE, 1.47–4.03°C; NSE, 0.54–0.90; d, 0.89–0.98; R2, 0.64–0.93) and moisture [water-filled porosity (wfps)] (MBE, 0.03–0.06 wfps; RMSE, 0.09–40.13 wfps; NSE, −0.24–0.49; d, 0.78–0.87; R2, 0.45–0.69) that agreed well with field measurements. Predicted daily soil CO2 fluxes (kg C ha−1) provided ‘good’ agreement with MBE (range −0.58−4.67), RMSE (range 2.10−7.36), d (range 0.68–0.93), NSE (range −0.92–0.79), and R2 (range 0.49–0.85). Statistics showed ‘poor’ agreement between the simulated and measured daily N2O emissions because peak emissions events in the measured data were less than predicted. Cumulative CO2 and N2O emissions and crop yields were well estimated by the model. Conclusions DNDCv.CAN simulated the impacts of diverse crop rotations and cover crops on soil moisture, temperature and greenhouse gas emissions in the humid south-east of USA. Implications Nitrogen transformation routines and effect of rainfall interception on soil water content need further investigation to address the variations in daily N2O emissions.